Glucose impairs B-1 cell function in diabetes

Authors

  • K. Jennbacken,

    1. Wallenberg Laboratory for Cardiovascular Research, Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author
  • S. Ståhlman,

    1. Department of Clinical Chemistry and Transfusion Medicine, Institute of Biomedicine, Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author
  • L. Grahnemo,

    1. Department of Rheumatology and Inflammation Research, Institute of Medicine, Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author
  • O. Wiklund,

    1. Wallenberg Laboratory for Cardiovascular Research, Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author
  • L. Fogelstrand

    Corresponding author
    1. Department of Clinical Chemistry and Transfusion Medicine, Institute of Biomedicine, Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden
    • Wallenberg Laboratory for Cardiovascular Research, Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy at University of Gothenburg, Gothenburg, Sweden
    Search for more papers by this author

Correspondence: L. Fogelstrand, Clinical Chemistry, Sahlgrenska University Hospital, Bruna Straket 16, SE-413 45 Gothenburg, Sweden.

E-mail: Linda.Fogelstrand@clinchem.gu.se

Summary

B-1 lymphocytes produce natural immunoglobulin (Ig)M, among which a large proportion is directed against apoptotic cells and altered self-antigens, such as modified low-density lipoprotein (LDL). Thereby, natural IgM maintains homeostasis in the body and is also protective against atherosclerosis. Diabetic patients have an increased risk of developing certain infections as well as atherosclerosis compared with healthy subjects, but the underlying reason is not known. The aim of this study was to investigate whether diabetes and insulin resistance affects B-1 lymphocytes and their production of natural IgM. We found that diabetic db/db mice had lower levels of peritoneal B-1a cells in the steady state-condition compared to controls. Also, activation of B-1 cells with the Toll-like receptor (TLR)-4 agonist Kdo2-Lipid A or immunization against Streptococcus pneumoniae led to a blunted IgM response in the diabetic db/db mice. In-vitro experiments with isolated B-1 cells showed that high concentrations of glucose, but not insulin or leptin, caused a reduced secretion of total IgM and copper-oxidized (CuOx)-LDL- and malondialdehyde (MDA)-LDL-specific IgM from B-1 cells in addition to a decreased differentiation into antibody-producing cells, proliferation arrest and increased apoptosis. These results suggest that metabolic regulation of B-1 cells is of importance for the understanding of the role of this cell type in life-style-related conditions.

Introduction

B-1 lymphocytes represent a unique innate-like cell population that can be subdivided into B-1a and B-1b based on presence or absence of CD5 on the surface [1]. They are distinguished from conventional adaptive B-2 cells by their surface phenotype, anatomical localization to peritoneal and pleural cavities, restricted use of VH genes that are minimally edited and their capacity for self-renewal. B-1 cells produce natural antibodies in a rapid T cell-independent manner in response to several microbial antigens [2, 3]. Natural antibodies, which in mice consist mainly of antibodies of the immunoglobulin (Ig)M isotype, are present at birth without external antigen stimuli and provide a first-line defence against invading microorganisms. Despite their overall weak binding properties and polyreactivity, they possess, together with complement, an important function in maintaining tissue homeostasis and clearance of apoptotic cells [4-6]. In both mice and humans, oxidation-specific epitopes found on altered self-antigens and apoptotic cells are dominant targets for natural antibodies [7]. In addition to B-1 cells, marginal zone B cells (MZB) in the spleen also contribute to the serum titres of natural IgM and they have functional properties in common with B-1 cells [8].

The regulation of B-1 cells is not understood completely, although both Toll-like receptor (TLR)-4 and TLR-2 agonists exert positive effects by inducing cell proliferation and secretion of natural antibodies [7]. In some conditions, B-1 cells and their antibodies seem to have protective properties while they are pathogenic in others. B-1 cells are increased markedly in number in autoimmune prone New Zealand black/New Zealand white (NZB/NZW) F1 mice, thereby linking these cells to autoimmunity [9]. Natural IgM promotes inflammation and tissue damage in several models of ischaemia–reperfusion injury [10, 11]. In contrast, B-1 cells and natural IgM have been assigned a protective role in atherosclerosis, which has been demonstrated in several in-vivo models [12-15]. In clinical studies, serum titres of IgM also correlate inversely with vascular risk [16-18]. The atheroprotective effect of natural IgM is proposed to be due to its binding to oxidized low-density lipoprotein (OxLDL), with the uptake of OxLDL being an important event in the development of atherosclerosis. Secreted IgM can bind to OxLDL in circulation or in the atherosclerotic plaque, thereby inhibiting the uptake of OxLDL by macrophage scavenger receptor, thus potentially decreasing foam cell formation [19, 20].

Individuals with diabetes have a several-fold increased risk of cardiovascular disease (CVD) compared with healthy subjects, but the underlying reason is not known. Decreased levels of IgM against a particle resembling OxLDL, malonedialdehyde-modified LDL (MDA-LDL) have been reported in individuals with diabetes [21-23]. Moreover, diabetes is associated with an increased risk of community-acquired pneumonia, a disease often caused by Streptococcus pneumoniae, for which our immune defence is highly dependent upon natural antibodies [24]. The aim of this study was to investigate whether diabetes and insulin resistance affect B-1 cells and their production of natural IgM. We found that diabetic db/db mice have lower levels of peritoneal B-1a cells and a decreased IgM response to pneumococcal immunization and TLR-4 activation. Furthermore, our in-vitro studies showed that glucose in high concentrations reduces B-1 cell IgM secretion and differentiation into antibody-producing cells concurrent with proliferation arrest and increased apoptosis.

Materials and methods

Mice husbandry and experimental procedures

Specific pathogen-free C57BL/6 mice were purchased from Taconic (Skensved, Denmark). For isolation of peritoneal B-1 cells, male and female C57BL/6 mice were fed a normal chow diet. As a model for insulin resistance, 8-week-old male C57BL/6 mice were assigned randomly to a low glycaemic control diet or a high-fat diet (Harlan Laboratories, Madison, WI, USA) for 12 weeks. On a caloric basis, the low glycaemic control diet contained 16·8% fat, 60·9% carbohydrate and 22·3% protein (3·3 Kcal/g), whereas the high-fat diet contained 60·3% fat, 21·3% carbohydrate and 18·4% protein (5·1 Kcal/g). The diets contained comparable amounts of vitamins and minerals. Male db/db mice and control mice (+/+ or +/db) on a C57BL/6 background from Jackson Laboratories (Bar Harbor, ME, USA), and db/db and wild-type controls (+/+) on a BKS background from Taconic, were maintained on a normal chow diet.

For in-vivo assessment of the effect of TLR-4 agonist, 10–12-week-old db/db mice (on a C57BL/6 background) and controls were injected intraperitoneally with 0·34 mg/kg of the TLR-4 agonist Kdo2-Lipid A (Avanti Polar Lipids, Inc., Alabaster, AL, USA) or vehicle. For immunization studies, 10–12-week-old db/db mice and controls (on a C57BL/6 or BKS background) and C57BL/6 mice maintained on diets for 3 months were injected intraperitoneally with 11·5 μg of a 23-valent vaccine (Pneumovax; Sanofi Pasteur MSD, Lyon, France), containing 0·5 μg each of 23 types of polysaccharides from S. pneumoniae or saline. As indicated for each experiment, body weight, plasma insulin, glucose and antibody titres were followed in longitudinal blood samples. Before blood sampling, mice were fasted for 4 h. Plasma glucose in blood samples from fasted, non-anaesthetized animals was determined with a glucose dehydrogenase method by using HemoCue® B-glucose microcuvettes (HemoCue®, Ängelholm, Sweden) and insulin was determined by a mouse insulin enzyme-linked immunosorbent assay (ELISA) (Mercodia, Uppsala, Sweden). Plasma triglycerides and cholesterol were measured using Konelab 20 Autoanalyzer (Thermo Electron Corporation, Vantaa, Finland). All mice were housed in a controlled environment and all experimental protocols were approved by the animal ethical committee in Gothenburg.

Isolation and culture of B-1 cells

For in-vitro experiments with B-1 cells, B-1a cells, B-1b cells and B-2 cells, peritoneal exudate cells (PECs) were harvested by peritoneal lavage from male and female C57BL/6 mice (aged 15–50 weeks) using ice-cold phosphate-buffered saline (PBS) supplemented with 0·5% heat-inactivated fetal calf serum (FCS) and 10 mmol/l ethylenediamine tetraacetic acid (EDTA). B-1 cells were isolated using flow cytometric cell sorting, as described previously [7]. Briefly, PECs were incubated with Fc block™ (BD Pharmingen, Franklin Lakes, NJ, USA) for 5 min at 4°C. For sorting of B-1 cells, this step was followed by staining with allophycocyanin (APC)-labelled anti-CD19 (clone 1D3), phycoerythrin (PE)-labelled anti-CD23 (clone B3B4) and fluorescein isothiocyanate (FITC)-labelled anti-CD3 (clone 17A2). For sorting of B-1a, B-1b and B-2 cells, Fc block incubation was followed by staining with FITC-labelled anti-CD23 (clone B3B4), PE-labelled anti-CD5 (clone 53-7·3) and APC-labelled anti-CD19 (clone 1D3) (all antibodies from BD Pharmingen). B cell populations were sorted using a fluorescence activated cell sorter (FACS) Aria II (BD Pharmingen) based on forward-scatter (FSC), side-scatter (SSC) and staining for CD3, CD5, CD19, CD23 as follows: B-1 cells: CD19+, CD3, CD23; B-1a cells: CD19+, CD23, CD5dim; B-1b cells: CD19+, CD23, CD5; and B-2 cells: CD19+, CD23+, CD5. Doublets were excluded using FSC-H, FSC-A. According to post-sort analysis, sorted B cell populations constituted >99% of all isolated cells. Isolated cells were seeded at 200 000 cells/ml in culture medium containing RPMI-1640 supplemented with 10% heat-inactivated FCS, 20 mmol/l HEPES, 2 mmol/l glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mmol/l sodium pyruvate, 1 mmol/l nonessential amino acids and 0·05 mmol/l 2-mercaptoetanol (all Invitrogen, Carlsbad, CA, USA). As indicated for each experiment, cells were cultured at 37°C/5% CO2 for 3 or 7 days in the presence of D-(+)-glucose (Sigma, St Louis, MO, USA) at the concentrations indicated (5·5, 25, 50 or 75 mmol/l), Kdo2-Lipid A (100 ng/ml) (Avanti Polar Lipids, Inc.), mannitol (75 mmol/l), insulin (200–10 000 pmol/l) or leptin (0·01–1 μg/ml). Cell counting was performed at the end of the culture using a Countess® Automated Cell Counter (Invitrogen, Life Technologies, Paisley, UK), according to the manufacturer's instructions.

Flow cytometry

For analyses of leucocyte populations in peritoneum and spleen, PEC were harvested as described above and splenocytes were collected on a mesh filter, using ice-cold PBS supplemented with 0·5% heat-inactivated FCS and 10 mmol/l EDTA. For cell surface staining, PECs, single cell splenocyte suspensions or cultured B-1 cells were incubated with Fc block™ (clone 2·4G2) for 5 min at 4°C, followed by staining for 30 min as follows. Peritoneal cells were stained with FITC-labelled anti-CD23 (clone B3B4), peridinin chlorophyll-cyanin 5·5 (PerCP-Cy5·5)-labelled anti-CD11b (clone M1/70), PE-labelled anti-CD5 (clone 53-7·3), APC-labelled anti-CD19 (clone 1D3) and PE-Cy7–labelled anti-IgM (clone R6-60·2). Splenocytes were stained with PE-labelled anti-CD25 (clone 3C7), FITC-labelled anti-CD4 (clone RM4-5), APC-labelled anti-CD8a (clone 53-6·7), PE-Cy5-labelled anti-CD3 (clone 17A2), PE-Cy7-labelled anti-CD127 (clone HIL-7R-M21), PE-Cy7-labelled anti-IgM (clone R6-60·2), PE-labelled anti-CD5 (clone 53-7·3), PerCP-Cy5·5-labeled anti-CD19 (clone 1D3), APC-labelled anti-CD21/35 (clone 7G6) and FITC-labelled anti-CD23 (clone B3B4). Cultured B-1 cells were stained with PE-labelled anti-CD138 antibody (clone 281-2) (all antibodies from BD Pharmingen). For assessment of proliferation, freshly isolated B-1 cells were stained with 2 μmol/l CellTrace™ CFSE (Invitrogen), according to the manufacturer's protocol, before the experiment. At the end of the experiment, cells were harvested and directly resuspended for analysis. For apoptosis assays, cultured B-1 cells were stained with FITC-labelled annexin V (FITC annexin V apoptosis detection kit; BD Pharmingen) and cell viability solution containing 7-aminoactinomycin D (7-AAD) (BD Pharmingen), according to the manufacturer's instructions. Cells were analysed using a FACS Aria II (BD Pharmingen) and at least 100 000 cells were counted per sample in in-vivo experiments and at least 5000 cells in in-vitro experiments, with dead cells excluded based on FSC.

Antibody measurements

Specific IgM and IgG antibodies were determined in plasma and in cell culture supernatants by chemiluminescent ELISA, as described previously [7]. For detection of total IgM, microtitre plates were coated with purified rat anti-mouse IgM (2 mg/l) (clone II/41; BD Pharmingen). For the analysis of specific IgMs, microtitre plates were coated with copper oxidized (CuOx)-LDL (5 mg/l), MDA-LDL (5 mg/l) or Pneumovax (10 mg/l). CuOx-LDL and MDA-LDL were prepared from human LDL, as described previously [25]. Plates were post-coated with Tris-buffered saline (TBS) containing 1% bovine serum albumin (BSA) and samples were incubated for 1 h. Serum samples were diluted in TBS containing 1% BSA to a final dilution of 1:300 for detection of total IgM, 1:100 for IgM against CuOx-LDL and MDA-LDL and 1:50 for IgM against Pneumovax. Cell culture supernatants were diluted 1:125 for detection of total IgM and IgM against CuOx-LDL and MDA-LDL. Antibodies in samples were detected with alkaline phosphatase-conjugated goat anti-mouse IgM (μ-chain specific; Sigma-Aldrich) and quantified with Lumiphos 530 (Lumigen, Inc., Southfield, MI, USA) using LMaxII (Molecular Devices, Sunnyvale, CA, USA).

RNA extraction, cDNA synthesis and real-time polymerase chain reaction (PCR)

Total RNA from isolated peritoneal B-1 cells and positive control tissue (mouse liver, skeletal muscle and placenta) was extracted with the RNeasy micro prep kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. For assessments of mRNA for glucose transporters and the insulin receptor, cDNA was prepared with the VILO Superscript synthesis kit (Invitrogen), according to the manufacturer's instructions, and gene expression was evaluated with TaqMan gene expression assays for glucose transporter type 1 (GLUT1), GLUT2, GLUT3, GLUT4, the insulin receptor and endogenous control 18S (all from Applied Biosystems, Applera Corporation, Foster City, CA, USA). Real-time reverse transcription–polymerase chain reaction (RT–PCR) was performed with the ABI 7900 HT (Applied Biosystems) and PCR parameters were analysed according to the manufacturer's protocol. Relative gene expression was calculated with the ΔΔCt method. PCR reactions for target genes and control were performed in triplicate for all samples.

Statistical analyses

All statistical analyses were performed using spss software package version18. Comparisons between two independent groups were performed using the Mann–Whitney U-test or Student's t-test. For cell culture experiments, statistical analyses were performed with one-way analysis of variance (anova) with Dunnett's T3 or Tukey's post-hoc tests. Data are presented as mean ± standard error of the mean (s.e.m.) and P < 0·05 was considered statistically significant.

Results

Diabetic mice have a lower proportion of B-1a cells in the peritoneal cavity

As a model for diabetes, we compared db/db mice with their lean controls. At 10 weeks of age, the db/db mice (on a C57BL/6 background) had increased body weight, elevated plasma glucose and insulin levels, moderately increased levels of cholesterol and similar levels of triglycerides compared with control mice (Fig. 1a–d). In order to investigate if diabetes influenced immune cell distributions, PECs and splenocytes were collected and analysed with FACS. In the peritoneal cavity, the absolute numbers of B cells, T cells, macrophages, B-1a, B-1b and B-2 were significantly higher in the db/db mice than in control mice (Table 1), which might reflect an increased body weight and surface area in the peritoneal cavity of the db/db mice. Strikingly, the proportion of B-1a cells, expressed as percentages of total B cells, was lower in the db/db mice compared with the controls. The fraction of B-1b cells was similar in db/db mice and controls and, consequently, peritoneal B-2 cells expressed as a percentage of total B cells were higher in the db/db mice than in controls (Fig. 2). There were no differences in percentages of follicular B cells, MZB or B-1 cells in the spleen (Table 1). In conclusion, these results show that at steady state, db/db mice have a lower proportion of B-1a cells in the peritoneal cavity.

Figure 1.

Metabolic parameters in the db/db mice. Db/db mice and their controls (+/+ or +/db), on a C57BL/6 background, were followed from 5 to 10 weeks of age, when they were used in immunization experiments. (a) Body weight over time. (b) Glucose concentration, (c) insulin concentration and (d) triglyceride and cholesterol concentrations in plasma, all at 10 weeks of age. Values are mean ± standard error of the mean. n = 16 in (a–c) and n = 8 in (d). ***P ≤ 0·001; *P < 0·05.

Figure 2.

B cells in peritoneal cavity in db/db mice and controls (+/+ or +/db). (a) Peritoneal B-1a, B-1b and B-2 cells as percentage of total B cells in controls and db/db mice. Representative fluorescence activated cell sorter (FACS) plots from (b) one control mouse and (c) one db/db mouse. Total peritoneal cells were gated on CD19 for B cells and subsequently B-1a cells were defined as CD5+, CD23, B-1b cells as CD5, CD23, and B-2 cells as CD5, CD23+. Values are mean ± standard error of the mean; n = 8. ***P ≤ 0·001.

Table 1. Distribution of leucocytes in peritoneum and spleen and circulating IgM levels
 Controldb/dbLFDHFD
  1. ***P ≤ 0·001; **P ≤ 0·01; *P < 0·05. Absolute cell numbers and percentages for db/db and controls and for C57BL/6 mice maintained on low glycaemic control diet (LFD) and high-fat diet (HFD) for 3 months (n = 8 in each group). Absolute cell numbers are presented as 105 for peritoneal cells. Values are mean ± standard error of the mean. Ig: immunoglobulin; LDL: low-density lipoprotein; CuOx: copper-oxidized; MDA: malondialdehyde; MZB: marginal zone B cells.
Peritoneal cavity
B cells (absolute number)1·8 ± 0·35·3 ± 1·0**2·6 ± 0·54·4 ± 0·7*
T cells (absolute number)0·3 ± 0·11·2 ± 0·3**0·4 ± 0·10·9 ± 0·1**
MQ (absolute number)2·0 ± 0·24·3 ± 0·6***1·9 ± 0·32·1 ± 0·4
B-1a (absolute number)1·1 ± 0·22·0 ± 0·4*1·1 ± 0·21·4 ± 0·2
B-1b (absolute number)0·5 ± 0·11·3 ± 0·2**0·8 ± 0·11·1 ± 0·1
B-2 (absolute number)0·2 ± 0·12·1 ± 0·4***0·7 ± 0·22·0 ± 0·3**
B cells (% of total)43 ± 1·848 ± 2·350 ± 5·560 ± 2·1
T cells (% of total)6·4 ± 0·810·6 ± 1·4**7·4 ± 0·912·6 ± 1·3**
Spleen
Spleen weight (mg)82 ± 1·586 ± 3·042·6 ± 0·360·7 ± 0·4**
B cells (% of total)61 ± 1·464 ± 1·054 ± 2·865 ± 1·1*
Follicular B (% of total B)85 ± 0·886 ± 0·679 ± 0·881 ± 1·0
MZB (% of total B)6·7 ± 0·57·1 ± 0·710·3 ± 0·46·0 ± 0·7***
B-1 (% of total B)3·7 ± 0·34·5 ± 0·42·5 ± 0·22·7 ± 0·3
T cells (% of total)39 ± 1·436 ± 1·046 ± 2·834 ± 1·1*
CD8+ T cells (% of total T)23 ± 1·225 ± 1·130 ± 1·023 ± 0·5***
CD4+ T cells (% of total T)49 ± 1·144 ± 1·3*43 ± 1·245 ± 1·0
IgM antibodies
Total IgM (a.u.)1·3 ± 0·11·5 ± 0·1*1·7 ± 0·21·8 ± 0·2
MDA-LDL IgM (a.u.)2·0 ± 0·22·7 ± 0·2**1·3 ± 0·21·6 ± 0·2
CuOx-LDL IgM (a.u.)3·1 ± 0·43·7 ± 0·62·0 ± 0·32·6 ± 0·5

In accordance with the overall increased absolute number of B cells in the db/db mice, the basal levels of total IgM and IgM against MDA-LDL were higher in db/db mice than control mice at 10 weeks of age (Table 1).

Lower IgM response in diabetic mice after TLR activation or immunization with S. pneumoniae

In order to investigate if the decreased proportion of B-1a cells in diabetic mice is reflected by a blunted innate humoral response, db/db mice and controls (on a C57BL/6 background) were injected intraperitoneally with the TLR-4 agonist Kdo2-Lipid A. As expected, injection of Kdo2-Lipid A induced an increase in IgM against CuOx-LDL and MDA-LDL in plasma in both diabetic and control mice. The IgM response was lower in the db/db mice than in control mice, both at 3 and 7 days post-injection (Fig. 3a and b). Increases in total IgM were not consistent, but at day 3 post-injection the response was lower in the db/db mice than in control mice, and after 1 week the increase in IgM in the db/db mice had reached the response observed in controls (Fig. 3c). This suggests that the innate immune system in db/db mice has a delayed and blunted response to bacterial components. Except for an increase in peritoneal B-1b cells in both db/db and controls, stimulation of TLR-4 did not result in significant changes in population sizes of subsets of B cells or T cells in spleen or the peritoneal cavity (data not shown).

Figure 3.

Immunization of db/db and controls (+/+ or +/db). At 10 weeks of age the db/db mice and controls were injected with (a–c) the Toll-like receptor (TLR)-4 agonist Kdo2-Lipid A or (d) Pneumovax. Blood samples were taken at the indicated time-points and immunoglobulin (Ig)M antibodies were assessed with enzyme-linked immunosorbent assay (ELISA). (a) IgM against copper-oxidized low-density lipoprotein (CuOx-LDL), (b) IgM against malondialdehyde (MDA)-LDL, (c) total IgM, (d) IgM against Pneumovax. Data are presented as IgM response for TLR-4 agonist/Pneumovax at day X divided with mean IgM response for vehicle group at day X. Values are mean ± standard error of the mean; n = 3–7 in (a–c) and n = 8 in (d). **P ≤ 0·01; *P < 0·05.

To explore further the effect of diabetes on the humoral innate response known to be exerted by B-1 cells, we immunized another set of db/db mice and controls with Pneumovax, a vaccine composed of 23 polysaccharides from S. pneumoniae. Upon immunization, the response to the vaccine, assessed as plasma IgM directed against Pneumovax, was blunted in the db/db mice compared with the control mice (Fig. 3d). The Pneumovax immunization did not result in significant changes in population of subsets of B cells and T cells in control mice or in diabetic mice (data not shown). We also performed the immunization experiment on a set of db/db mice on BKS background and BKS controls. These db/db animals showed more severe diabetes with higher plasma glucose levels and low insulin levels (compared with the db/db on a C57BL/6 background). The response to Pneumovax immunization at 7 days was blunted in the db/db mice (the IgM directed against Pneumovax response in db/db was 61% ± 3·3 of the response in controls). Together, these experiments show that diabetic mice have a dampened response to stimuli that require a functional humoral innate immune response.

No effect of insulin resistance on IgM response to immunization with S. pneumoniae

In order to compare the results obtained in the db/db mice on a C57BL/6 background, which are all diabetic and insulin-resistant, with mice that were insulin-resistant but not overtly diabetic, we performed experiments on C57BL/6 mice in which we induced insulin resistance with a high-fat diet. Mice were fed either a high-fat diet, based on lard, or a low glycaemic control diet for 3 months. At the end of this period, mice on the high-fat diet had significantly increased body weight and insulin levels (Fig. 4a and b), but they showed only moderately increased plasma glucose (14·5 mmol/l ± 0·48 versus 11·2 mmol/l ± 0·25, P ≤ 0·001), triglycerides (2·1 mmol/l ± 0·09 versus 1·3 mmol/l ± 0·06, P ≤ 0·001) and total cholesterol (5·9 mmol/l ± 0·28 versus 2·6 mmol/l ± 0·16, P ≤ 0·001) compared with mice receiving the control diet. Similar to the db/db mice, mice on the high-fat diet showed decreased proportions of B-1a cells, expressed as a percentage of total B cells, and also of B-1b cells, compared with the mice receiving control diet. There was also a corresponding increase in the proportion of B-2 cells (Fig. 4c). Compared with the results in the db/db mice, the differences in absolute numbers of the different cell populations in the peritoneal cavity were not as obvious (Table 1). In the spleen, the numbers of MZB cells, expressed as a percentage of total B cells, were significantly lower in mice on the high-fat diet (Table 1). There were no significant differences in the plasma levels of total IgM or IgM against CuOx-LDL and MDA-LDL between mice on the high-fat and control diets (Table 1). To assess the humoral innate response, mice that had been on the diets for 3 months were immunized with Pneumovax. The IgM response was similar to the response in control mice on C57BL/6 mice used in the immunization experiment with db/db mice. Although there was a slightly delayed response in the mice on the control diet, there were no differences between the two diets at 7 days after immunization (Fig. 4d), nor were there any differences in subsets of B or T cells in the spleen or in the peritoneal cavity between mice immunized with vehicle or Pneumovax (data not shown). Together with the results in db/db mice, these findings indicate that diabetes, but not insulin resistance, is associated with a blunted humoral innate response.

Figure 4.

Insulin-resistant model. C57BL/6 mice were fed a high-fat diet or a low glycaemic control diet. (a) Body weight over time, (b) insulin levels after 3 months on diet, (c) B-1a, B-1b and B-2 cells in peritoneum displayed as percentage of total B cells after 3 months on diet. Total peritoneal cells were gated on CD19 and B-1a cells defined as CD5+, CD23, B-1b cells defined as CD5, CD23 and B-2 cells defined as CD5, CD23+. (d) Response to immunization. At 10 weeks of age the mice on high-fat diets and controls were immunized with Pneumovax or vehicle. Blood samples were taken at the indicated time-points and IgM antibodies against Pneumovax were assessed with enzyme-linked immunosorbent assay (ELISA). Data are presented as IgM response for Pneumovax at day X divided with mean IgM response for vehicle group at day X. Values are mean ± standard error of the mean; n = 16 in (a,b) and n = 8 in (c,d). **P ≤ 0·01; *P < 0·05. HFD: high-fat diet; LFD: low-fat diet.

High glucose concentrations inhibit IgM secretion from B-1 cells in vitro

Because diabetes seemed to influence the function of B-1 cells in db/db mice, we continued to investigate the effects of metabolic factors on B-1 cells, B-1a cells, B-1b cells and B-2 cells in vitro, using FACS-purified mouse peritoneal B cell subpopulations from C57BL/6 mice. Isolated B-1 cells were cultured in the presence of increasing concentrations of glucose, insulin or leptin. As we have shown earlier, cultured B-1 cells secrete low levels of IgM, and addition of a TLR agonist results in a robust increase in the release of IgM [7]. As shown in Fig. 5a, stimulation of TLR-4 with Kdo2-Lipid A induced substantially the secretion of total as well as anti-CuOx-LDL and anti-MDA-LDL IgM, but this induction was gradually diminished in the presence of increasing concentrations of glucose. When IgM levels in the supernatants were related to B-1 cell numbers there was still a trend, although not statistically significant, towards a negative effect of glucose for IgM against CuOx-LDL and MDA-LDL (Fig. 5b). Secretion of IgM against CuOx-LDL and MDA-LDL was also investigated in B-1a, B-1b and B-2 populations separately. As shown in Fig. 5c and d, all three cell types produced IgM directed against CuOx-LDL and MDA-LDL upon TLR stimulation. This IgM secretion was inhibited by glucose in all three cell types, shown most consistently in B-1a cells (Fig. 5c and d), and accompanied by decreased cell numbers (data not shown). There was no effect of an equal concentration of mannitol, ruling out the possibility that the effect of glucose was due to osmotic stress (Fig. 5a–d). Culture of B-1 cells in the presence of increasing concentrations of insulin or leptin did not affect TLR-4-induced IgM secretion (data not shown). Together, these results indicate that high glucose concentrations have a negative impact on the activation of B-1 cells.

Figure 5.

Effects of glucose on immunoglobulin (Ig)M secretion, proliferation, apoptosis, death and differentiation into antibody-producing cells in isolated mouse peritoneal B cells. Peritoneal B-1 cells (a–b, e–h) and B-1a, B-1b and B-2 cells (c–d) were isolated with fluorescence activated cell sorter (FACS) and cultured in the absence or presence of Toll-like receptor (TLR)-4 agonist Kdo2-Lipid A. Culture medium contained different concentrations of glucose as indicated and mannitol was included as an osmotic control. After 3 days, cells were analysed with FACS, and after 7 days IgM antibodies were determined with enzyme-linked immunosorbent assay (ELISA). (a) Total IgM and IgM against copper-oxidized low-density lipoprotein (CuOx-LDL) and malondialdehyde (MDA)-LDL from B-1 cells. (b) Total IgM and IgM against CuOx-LDL and MDA-LDL from B-1 cells related to cell number, as assessed by counting of viable cells at the end of the culture. (c) IgM against CuOx-LDL from B-1a, B-1b and B-2 cells. (d) IgM against MDA-LDL from B-1a, B-1b and B-2 cells. (e) Proliferation of B-1 cells based on staining with carboxyfluorescein succinimidyl ester (CFSE) at culture start. (f) Apoptosis and (g) cell death in B-1 cells evaluated with annexin V and 7-aminoactinomycin D (7-AAD), respectively. (h) Differentiation of B-1 cells into antibody-producing cells based on CD138-positive cells. Values are mean ± standard error of the mean; n = 2–4. Glc: glucose; concentrations are in mmol/l; Mnt: mannitol, concentration in mmol/l; a/*** P ≤ 0·001, b/** P ≤ 0·01, c/*, P < 0·05.

High glucose concentrations result in proliferation arrest and increased apoptosis

In order to investigate possible mechanisms for the effects of glucose on IgM production, proliferation, apoptosis and differentiation were determined in isolated B-1 cells cultured in the presence of increasing concentrations of glucose. Proliferation was assessed by staining cells with CFSE before the start of the culture, followed by FACS analysis at harvest. Percentage of cells undergoing proliferation decreased from 70% at physiological glucose concentration to 40% at 75 mmol/l glucose (Fig. 5e). We also analysed the percentage of apoptotic and dead cells (late apoptotic) in B-1 cell cultures by using staining with annexin V in combination with 7-AAD. With increasing glucose concentrations, both the proportion of apoptotic and dead cells increased (Fig. 5f and g). In unstimulated cells (cells cultured in the absence of TLR-4 agonist), the proportion of dead cells was the highest. As a marker for differentiation into an antibody-producing cell, cultured B-1 cells were stained for the plasma cell marker CD138. Upon TLR-4 stimulation, approximately 35% of cells expressed CD138, compared with approximately 18% among the unstimulated cells. Increasing concentrations of glucose resulted in a decreased percentage of CD138-expressing cells (Fig. 5h), indicating that fewer cells differentiated to IgM-secretion. Mannitol, in a concentration corresponding to the highest glucose concentration, did not affect proliferation, apoptosis or CD138 expression (Fig. 5e–h). As interleukin (IL)-10 has been shown previously to affect proliferation of B-1 cells [26], we assessed the levels of this cytokine in the medium at the end of the culture. Levels of IL-10 in were not affected by glucose concentration (IL-10 levels in 25, 50 and 75 mmol/l glucose relative to 5·5 mmol/l were 81% ± 8·8, 105% ± 23·6 and 67% ± 13·5, respectively).

Isolated B-1 cells express GLUT1, GLUT2 and the insulin receptor

Because high glucose concentrations, but not insulin, affected B-1 cell function in our experiments, we investigated the mRNA expression of glucose transporters and the insulin receptor in isolated B-1 cells. Peritoneal B-1 cells expressed mRNA encoding for GLUT1 (2−ΔΔCt = 0·05 ± 0·002 relative placenta), GLUT3 (2−ΔΔCt = 0·34 ± 0·002 relative placenta) and the insulin receptor (2−ΔΔCt = 0·65 ± 0·04 relative skeletal muscle) but not mRNA encoding for GLUT2 or GLUT4 (undetectable levels, positive control tissue were liver and skeletal muscle, respectively).

Discussion

Components of the immune system are disturbed in diabetes. The immunological changes include altered numbers and activation states of various leucocyte populations and changes in specific cytokines and chemokines [27], and it is well known that diabetes is associated with several infections [28]. For example, diabetes is associated with an increased risk of community-acquired pneumonia, a disease often caused by S. pneumoniae, for which our immune defence is highly dependent upon the innate immune system [24]. In line with this, it has been shown that titres of IgM antibodies against MDA-LDL are decreased in individuals with diabetes [21-23]. Also, levels of IgM against phosphorylcholine show an inverse relationship with body mass index (BMI) and waist circumference and has been associated with a western lifestyle [29]. In the present study we found that at steady state, diabetic db/db mice have lower proportions of B-1a cells in the peritoneal cavity. The db/db mice also showed a dampened antibody response when their innate immune system was challenged with a TLR-4 ligand or pneumococcal components, indicating that the B-1 cells in the db/db mice were less responsive in producing protective IgM. In accordance with this, decreased IgM production in response to LPS treatment has been reported previously in a mouse model of type I diabetes [30]. Together, these results indicate that diabetes suppresses innate immune responses challenged with T independent antigens, at least in mice. This inhibitory effect of glucose at high concentrations is not necessarily specific for B-1a or B-1b cells, as supported by our in-vitro findings in sorted B cell subpopulations. The decreased proportion of B-1a cells in the peritoneal cavity of db/db mice was not accompanied with decreased IgM levels at steady state. However, previous studies have shown that B-1 cells in pleural and peritoneal cavities secrete only small amounts of natural antibodies at steady state [31], which corresponds with their low levels of mRNA encoding secreted IgM [32]. Instead, it seems that spleen and bone marrow contain B-1 cells that secrete spontaneously large amounts of IgM that are thought to be a major contributor to circulating levels of IgM [31]. The decrease in proportion of B-1a cells in the diabetic mice was accompanied by an increase in B-2 cells. Therefore, we cannot rule out that the proportion of B-1a cells might be influenced by the high number of B-2 cells. The reason for a concomitant increase in B-2 cells is unclear. By performing in-vitro experiments with isolated B-2 cells, where glucose also had an inhibitory effect on this cell type, we conclude that the high number of B-2 cells in the diabetic mice is not a direct effect of glucose. Hypothetically, there might be a higher antigenic burden in these mice due to an overall effect on the innate immune system.

Hyperglycaemia is one of the key factors that contribute to diabetic complications. Prolonged exposure to high glucose have many effects, including release of reactive oxygen species (ROS) and several proinflammatory cytokines [33-35], and therefore have deleterious effects on cells and cellular processes. Here we found that hyperglycaemia affected isolated mouse peritoneal B-1 cells and the production of IgM. Increasing concentrations of glucose resulted in diminished secretion of total IgM and IgM against CuOx-LDL and MDA-LDL. We also found that a high glucose concentration increased apoptosis and cell death and affected the proportion of cells in mitosis in the B-1 cells negatively. Our results are in line with previous reports showing that chronic exposure to hyperglycaemia increases apoptosis, an effect mediated via impaired phosphorylation of the protein kinase Akt [36, 37]. Also, increased apoptosis, together with ROS production and lipid peroxidation, has been observed in B lymphocytes isolated from diabetic mice [30]. In addition to affecting apoptosis, high glucose affects cellular survival and proliferation progressively. For example, exposure of T and B lymphocytes to high glucose results in inhibition of DNA synthesis and proliferation [30, 38].

B cells, together with other immune cells, are implicated in the pathogenesis and progression of atherosclerosis. Diabetic patients have an increased risk of developing atherosclerosis, and a disturbed function of B-1 cells as shown in this study could possibly mediate this. Previous studies have suggested that B-1a cells and natural IgM are atheroprotective [15], probably by the ability of these antibodies to compete with macrophages in binding OxLDL, thereby inhibiting foam cell formation [19]. In mice, absence of IgM leads to an increased propensity for atherosclerosis [12] and atherosclerosis development is inhibited if the amount of oxidation-specific epitopes is increased, such as after immunization with the bacteria S. pneumoniae [13]. Clinical studies have shown that elevated circulating levels of IgM against OxLDL are associated with reduced vascular risk in humans, but IgG antibodies show variable associations [16-18].

In conclusion, this study shows that diabetic db/db mice have lower proportion of peritoneal B-1a cells in the steady state and show a dampened response to TLR activation and immunization against S. pneumoniae, both stimuli that require a functional innate immune system. Moreover, culture of isolated peritoneal mouse B-1 cells in high glucose concentrations led to reduced IgM secretion, decreased proliferation, and increased apoptosis. The results suggest that metabolic regulation of B-1 cells is of importance for the understanding of the role of this cell type in lifestyle-related conditions.

Acknowledgements

This study was supported by the Swedish Heart and Lung Foundation, the Swedish Research Council, Sahlgrenska University Hospital, the Swedish Society of Medicine, the research foundations of Åke Wiberg, Syskonen Svensson, Fredrik and Ingrid Thuring, Magnus Bergvall and the Emelle Foundation. We thank Hannah Shaffer for excellent laboratory assistance.

Disclosure

The authors declare no conflict of interest.

Ancillary